JP7346738B2 - Cobalt chemistry for smooth topology - Google Patents

Description

The compositions and processes described herein generally relate to electrolytic deposition chemistry and methods for depositing cobalt and cobalt alloys. These compositions and methods are used for cobalt-based metallization of interconnect features in semiconductor substrates. Screening methods for identifying suitable bath additives are also described.

In damascene processing, electrical interconnections are formed within an integrated circuit substrate by filling interconnect features such as vias and trenches formed within the substrate with metal. Copper is the preferred conductor for electronic circuits. Unfortunately, when copper is deposited on a silicon substrate, it can rapidly diffuse into both the substrate and the dielectric film (such as SiO ₂ or a low-k dielectric). Copper also has a tendency to migrate from one location to another and form voids and ridges when electrical current passes through active interconnect features. Copper can also be diffused into device layers built on top of the substrate in multilayer device applications. Such diffusion can be harmful to the device because it can damage adjacent interconnects and/or cause electrical leakage between two interconnects. Electrical leakage can result in electrical shorts, and corresponding diffusion from interconnect features can disrupt electrical flow.

With the reduction in size in electronic devices and the increase in desired performance in electronic devices, the need for defect-free, low resistivity interconnects has become important in the electronic packaging industry. As the density of integrated circuits within microelectronic devices continues to increase with each generation or node, interconnects become smaller and their aspect ratios generally increase. The build-up process using barrier and seed layers, prior to damascene copper electroplating, currently has drawbacks that become more apparent as the demand for higher aspect ratio features and higher quality electronic devices increases. So I'm having trouble. As a result, more suitable plating chemistries are needed to enable defect-free metallization.

When submicron vias and trenches are filled by electrolytic deposition of copper, a barrier layer is first applied over the cavity walls to prevent copper diffusion and electromigration into the surrounding silicon or dielectric structure. It is generally necessary to precipitate it. A seed layer is deposited on top of the barrier layer to establish a cathode for electrodeposition. Barrier and seed layers can be very thin. This can be diluted, especially if the electroplating solution contains a suitable blend of accelerators, inhibitors, and leveling agents. However, as electronic circuit densities continue to increase and via and trench entrance dimensions become smaller, even very thin barrier and seed layers will occupy a larger portion of the entrance dimensions. As the pores reach dimensions of less than 50 nm, especially less than 40 nm, less than 30 nm, less than 20 nm, or even less than 10 nm (8 or 9 nm), it becomes increasingly difficult to fill the cavities with copper deposits completely free of voids and seams. become. The most advanced features have a bottom width of only 2-3 nm, a center width of about 4 nm, and a depth of 100-200 nm, which translates into aspect ratios of about 25:1 to about 50:1. It means that it has.

Electrolytic deposition of cobalt is performed in a variety of applications in the manufacture of microelectronic devices. For example, cobalt is used in capping damascene copper metallization used to form electrical interconnects within integrated circuit boards. However, since cobalt deposits have a higher resistivity, such processes do not offer a good alternative to copper electrodeposition in via or trench filling to provide the primary interconnect structure. . In typical semiconductor processing, a chemical mechanical polishing/planarization (CMP) step is performed after electrodeposition to polish away overplated deposits or excess amounts. It is important to have a smooth topology for overloading, as rough or uneven surfaces can cause CMP defects. Additionally, rough or uneven surfaces can be particularly problematic due to differences in over-the-top (OB) thickness between features and non-feature areas, which is typically expected due to kinetic plating. These differences in excess thickness can be further enhanced with feature density.

US Pat. The method and equipment are described. Doubina uses a variety of plating additives, including combinations of promoters and inhibitors and specific conductivities, to achieve the desired plating effects. However, Doubina does not mention that it is desirable to minimize excessive thickness changes between feature areas and non-feature areas of the substrate, nor does Doubina mention that it is desirable to minimize excessive thickness changes between feature areas and non-feature areas of the substrate, nor does he mention that it is desirable to There is also no mention of any method of evaluating the effectiveness of the bath additives therein.

U.S. Patent Publication No. 2009/0188805 by Moffat et al., incorporated herein by reference in its entirety, the subject matter of which is the use of at least one ferromagnetic material in a substrate, which may be nickel, cobalt, or iron. This paper describes a method of electrodeposition in a three-dimensional pattern. Although Moffat describes controlling the potential between an electrode and a counter electrode, he does not describe any method for evaluating the effectiveness of bath additives.

US Pat. Compositions for and methods of using such compositions are described. The interconnect features are metallized by contacting the semiconductor-based structure with a cobalt electrolytic composition, and an electrical current is supplied to the electrolytic composition to deposit cobalt onto the base structure and decouple the submicron-sized features with cobalt. Fill with.

US Pat. ing. However, Josell's process involves additives such as the so-called "coin metals" of gold, silver, and copper, rather than cobalt plating baths. Although Josell describes electrochemical measurements, Josell does not mention that it is desirable to minimize excessive thickness changes between feature regions and non-feature regions of the substrate; Josell also does not contemplate how to evaluate the effectiveness of various bath additives.

It therefore proves desirable to provide a method for evaluating bath additives for use in cobalt electroplating baths to produce desirable results.

To that end, the inventors of the present invention have surprisingly used cyclic voltammetry to analyze and/or screen additives for use in cobalt electroplating compositions and to obtain suitable/desired results. It has been discovered that it is possible to determine additives that can produce .

Cyclic voltammetry is used to study qualitative information about electrochemical processes under various conditions, elucidating the electronic stoichiometry of the system, analyte diffusion coefficients, and formal reductions that can be used as identification tools. potential can be determined. Additionally, CV may be used to determine the concentration of an unknown solution by generating a current versus concentration calibration curve.

Cyclic voltammetry (CV) is a potentiodynamic electrochemical measurement method that can be used to probe reactions involving electron transfer. In cyclic voltammetry experiments, the working electrode potential is ramped linearly with time in the cyclic phase. During the initial forward scan (t ₀ -t ₁ ), a gradually reduced potential is applied. Therefore, the cathode current will increase over this period, assuming there is reducing analyte in the system, at least initially. At some point after the analyte reduction potential is reached, the cathodic current decreases as the concentration of reducing analyte decreases. If the redox bond is reversible during the reverse scan (t ₁ -t ₂ ), the reduced analyte begins to be reoxidized, generating a current of opposite polarity (anodic current) than before. The more reversible the redox bond, the more the oxidation peak has a shape similar to the reduction peak. Therefore, cyclic voltammetry can provide information regarding redox potential and electrochemical reaction rates.

The rate of voltage change over time during each of these steps is known as the experimental scan rate (V/sec). Potential is measured between the working and reference electrodes, and current is measured between the working and counter electrodes. These data are plotted as current (i) versus applied potential (E, often referred to simply as "potential").

After reaching the set potential in a CV experiment, the potential of the working electrode is ramped in the opposite direction back to the initial potential. These cycles of potential ramps may be repeated as many times as desired.

The aim of the present invention is to provide a cobalt precipitate exhibiting smooth topology against excessive thickness.

Another object of the invention is to minimize excessive thickness changes between feature and non-feature regions.

Another object of the invention is to provide a void-free cobalt deposit.

Yet another object of the present invention is to minimize impurities in cobalt electrodeposit.

Yet another object of the present invention is to provide a screening method for evaluating the effectiveness of various bath additives for use in cobalt electroplating baths.

To that end, in one embodiment, the present invention generally relates to an electroplated cobalt deposit on a surface of a substrate, wherein the electroplated cobalt deposit has a thickness over substantially the entire surface of the substrate of less than about 200 nm. Indicates the difference in strength.

In another embodiment, the present invention generally relates to a method of electrodepositing cobalt onto a substrate comprising recessed features and non-feature areas, wherein the cobalt deposit has a thickness over substantially the entire surface of the substrate of less than about 200 nm. This method shows the difference in
a) receiving a substrate within an electroplating chamber;
b) immersing the substrate in a cobalt electrolyte, the cobalt electrolyte comprising:
(1) A cobalt ion source,
(2) boric acid;
(3) a pH adjuster;
(4) immersing the substrate in a cobalt electrolyte, the organic additive comprising an inhibitor;
c) electroplating cobalt into the feature and onto the non-feature areas for a period of time and plating onto the non-feature areas under conditions that achieve horizontal, seamless bottom-up filling in the concave feature; and, including.

3 shows cyclic voltammetry curves of Comparative Example 1, Example 2, and Example 3 according to the present invention.

1 shows a photograph of cobalt precipitates described in Comparative Example 1.

Figure 2 shows a photograph of the cobalt precipitate described in Example 2.

A photograph of the cobalt precipitate described in Example 3 is shown.

As used herein, "a," "an," and "the" refer to the singular and plural unless the context clearly dictates otherwise. means both referents.

As used herein, the term "about" means a measurable value, such as a parameter, amount, duration, etc., of +/-15 of and from the specifically recited value. % or less, preferably +/-10% or less, more preferably +/-5% or less, even more preferably +/-1% or less, and still even more preferably +/- Variations of 0.1% or less are meant to be included, so long as such variations are appropriate for practicing the invention described herein. Furthermore, it is to be understood that the values meant by the modifier "about" are themselves specifically disclosed herein.

As used herein, spatial terms such as "beneath", "below", "lower", "above", "upper", etc. Relative terminology is used to facilitate explanation to describe the relationship of one element or feature to another element or feature(s) as shown in the figures. . It is further understood that the terms "front" and "back" are not intended to be limiting, but are intended to be interchangeable where appropriate.

As used herein, the terms "comprises" and/or "comprising" specify the presence of the described feature, integer, step, act, element, and/or component. but does not exclude the presence or addition of one or more other features, integers, steps, acts, elements, components, and/or groups thereof.

As used herein, "substantially free" or "essentially free" with respect to a particular element or compound means that the given element or compound is means not detected by conventional analytical means well known to those skilled in the art of metal plating for analysis. Such methods typically include atomic absorption spectroscopy, titration, UV-Vis spectroscopy, secondary ion mass spectrometry, and other commonly available analytical methods.

In one embodiment, the invention generally relates to an electroplated cobalt deposit on a surface of a substrate, wherein the electroplated cobalt deposit has a thickness difference over substantially the entire surface of the substrate of less than about 200 nm. shows.

In another embodiment, the present invention generally relates to a method of electrodepositing cobalt onto a substrate comprising recessed features and non-feature areas, wherein the cobalt deposit has a thickness over substantially the entire surface of the substrate of less than about 200 nm. This method shows the difference in
a) receiving a substrate within an electroplating chamber;
b) immersing the substrate in a cobalt electrolyte, the cobalt electrolyte comprising:
(1) A cobalt ion source,
(2) boric acid;
(3) a pH adjuster;
(4) an organic additive containing an inhibitor;
(5) immersing the substrate in a cobalt electrolyte, optionally including one or more additional bath additives;
c) electroplating cobalt into the feature and onto the non-feature areas for a period of time and plating onto the non-feature areas under conditions that achieve horizontal, seamless bottom-up filling in the concave feature; and, including.

The cobalt ion source may be selected from the group consisting of cobalt sulfate, cobalt chloride, sulfamide chloride, and combinations of one or more of the foregoing. In one embodiment, the cobalt ion source is cobalt sulfate heptahydrate.

In preferred embodiments, the cobalt electroplating composition is at least substantially free of other metal ions, including less than about 1% by weight, more preferably 0.1% by weight of other metal ions. %, most preferably less than about 0.01% by weight. Most preferably, the cobalt electroplating composition is free of any metal ions other than cobalt ions.

Moreover, the cobalt electroplating composition is preferably at least substantially free of copper ions. Although very trace copper contamination can be difficult to avoid, it is particularly preferred that the copper ion content of the bath is below 20 ppb, for example in the range 0.1 ppb to 20 ppb. For compositions defined herein, "substantially free of copper ions" means that less than 20 ppb copper ions are present in solution.

The cobalt ion concentration in the electroplating solution typically ranges from about 1 to about 50 g/L, preferably from about 2 to about 25 g/L, more preferably from about 2 to about 10 g/L, and more preferably from about 2 to about 25 g/L. It is in the range of about 5 g/L.

Preferably, the electrolytic cobalt composition also optionally includes a buffer to stabilize the pH. One preferred buffering agent is boric acid (H ₃ BO ₃ ), which can be incorporated into the composition at a concentration of about 5 to about 50 g/L, preferably about 15 to about 40 g/L. . The pH of the composition is preferably maintained within the range of about 0.5 to about 8. In one embodiment, the pH of the cobalt composition is preferably less than 5, more preferably in the range of about 1 to about 5, more preferably in the range of about 2 to about 4.5, most preferably in the range of about 2.5 to about 5. It is maintained within a range of about 4.

The composition may include one or more inhibitory compounds. In one embodiment, the one or more inhibitory compounds include an acetylene alcohol compound or derivative thereof, as further described herein. The concentration of the inhibitor is preferably from about 1 to about 500 mg/L, more preferably from about 5 to about 200 mg/L, most preferably from about 1 mg/L to about 70 mg/L, and most preferably from about 20 to about 50 mg/L. It is.

The composition may also optionally include one or more uniformity-enhancing compounds, preferably including amine polyol compounds or derivatives thereof. A preferred uniformity improver is ethoxylated propoxylated triisopropanolamine. In one embodiment, the uniformity enhancer has a molecular weight of about 5000 g/mol. Other preferred uniformity-enhancing compounds include ethoxylated propoxylated ethylenediamine, ethoxylated propoxylated diethylenetriamine, and ethoxylated propoxylated triethylenetetramine. If used, the concentration of the uniformity enhancer is preferably from about 10 to about 4000 mg/L, more preferably from about 100 to about 2000 mg/L, and most preferably from about 250 to about 1000 mg/L.

The composition may also optionally include one or more depolarizing compounds. In one embodiment, the one or more depolarizing compounds include terminally unsaturated compounds or derivatives thereof that are capable of depolarizing the plating potential. In one embodiment, the depolarizing compound may be selected from the group consisting of sodium propargyl sulfonate, acetylene dicarboxylic acid, acrylic acid, propiolic acid, vinyl phosphonate, and mixtures thereof. One preferred depolarizing compound is sodium propargyl sulfonate. If used, the concentration of depolarizing compound is preferably about 0.1 to about 5000 mg/L, more preferably about 10 to about 1000 mg/L, and most preferably about 100 to about 500 mg/L.

In one embodiment, the cobalt electrolyte composition is essentially free of chloride ions. This means that the chloride content is less than about 1 ppm, more preferably less than 0.1 ppm.

Preferably, the electroplating composition does not include any functional concentration of reducing agent effective to reduce cobalt ions (Co ²⁺ ) to metallic cobalt (Co ⁰ ). "Functional Concentration" means any concentration of a reducing agent that is either effective to reduce cobalt ions in the absence of an electrolytic current or that is activated by an electrolytic field to react with cobalt ions. means.

Furthermore, the electroplating composition is preferably at least essentially free of dispersed particles, which are macroscopic particulate solids in solution that are dispersed and can negatively interfere with the metal electroplating process. It means no or substantially no.

In another preferred embodiment, the cobalt composition also optionally, but preferably, includes one or more levelers, one or more accelerators, and/or one or more wetting agents. In other preferred embodiments, the cobalt composition is free, preferably at least substantially free, of promoters or depolarizers. In other preferred embodiments, the cobalt composition is free, preferably at least substantially free, of leveler.

When divalent sulfur compounds are removed from the plating bath, the sulfur content of the cobalt deposit is reduced, which has a beneficial effect on chemical mechanical polishing and circuit performance.

When the concentration of divalent sulfur in the plating solution is 1 mg/L or less, the electrolytic composition is substantially free of divalent sulfur compounds. Preferably, the concentration of the compound containing divalent sulfur atoms is 0.1 mg/L or less. Even more preferably, the concentration of divalent sulfur atoms is below the level of detection using analytical techniques common to those skilled in the art of metal plating.

To reduce internal stress in the cobalt precipitate, the electrolytic composition may include a stress reducing agent such as saccharin. When used, saccharin is present in the electrolytic composition at a concentration of about 10 to about 300 ppm, more preferably about 100 to about 200 ppm.

When the plating bath contains an inhibitor as described herein and optionally a uniformity enhancer, the superfilling process proceeds satisfactorily without the need for an accelerator. The suppressor in this invention helps drive current through the machine to make bottom-up filling an effective feature, and the uniformity-enhancing additive helps improve deposit uniformity. The composition is substantially free of reducing agents that reduce Co ²⁺ to Co ⁰ , divalent sulfur, copper ions, nickel ions, and iron ions.

It has also been discovered that certain depolarizing compounds can function in conjunction with suppressor compounds as described herein. These compounds depolarize the plating potential to efficiently plate interconnect features.

In one embodiment, the inhibitor is an acetylene inhibitor. The acetylene inhibitor preferably comprises an alkoxylated propargyl alcohol or a reaction product of propargyl alcohol and a second component. Examples of suitable acetylene inhibitors include, but are not limited to, alkoxylated propargyl alcohols or reaction products of propargyl alcohol with glycidol, propylene oxide, glycidol and propylene oxide, or propylene glycol and glycidol. In one embodiment, the alkoxylated propargyl alcohol is ethoxylated propargyl alcohol. However, other alkoxylated propargyl alcohols are also contemplated for use in the compositions described herein. Examples of initiators and reactive species for preparing acetylene inhibitors according to the invention are shown in Table 1 below.

In one embodiment, x and y are 0-20, more preferably 0-10. In another preferred embodiment, one of x or y is at least 1. Examples of preferred ratios include, but are not limited to:

x is 0, y is 1 to 3,

y is 0, x is 1 to 7,

x is 1-4 and y is 1-4.

Other ratios of x and y are also well known to those skilled in the art and can be used in the present invention.

Table 2 provides several examples of preferred acetylene inhibitors according to the present invention and formulated using the initiators and reactive species described herein.

In one embodiment, n is 0-20, more preferably 0-10, most preferably 1-7.

The electrolytic compositions described herein can be used in a method for filling submicron features of semiconductor-based structures A, including submicron electrical interconnect features in the bottom, sidewalls, and top openings. has a department. The submicron features comprise cavities within the base structure that are superfilled by rapid bottom-up precipitation of cobalt. A metallized substrate containing a seed conductive layer is formed on the interior surface of the submicron feature, for example by physical vapor deposition of a metal seed layer, preferably a cobalt metal seed layer, or by deposition of a thin conductive polymer layer. A metallized substrate is applied to the bottom and sidewalls, typically in the area surrounding the feature. A metallized substrate within the feature is contacted with an electrolytic composition and an electrical current is applied to the electrolytic composition to cause electrodeposition of cobalt that fills the submicron feature. Through the joint action of the inhibitor, any uniformity enhancer, and any depolarizing compound, a vertical polarization gradient is created in features where filling occurs by bottom-up precipitation with vertical growth rates greater than horizontal growth rates. The resulting cobalt interconnect is substantially free of voids and other defects.

Bottom-up filling means that the precipitate grows from the bottom of the feature, such as a trench. Bottom-up filling can be classified as V-shaped or U-shaped. A V-shaped bottom-up has a pointed bottom and a U-shaped bottom-up has a flatter bottom. U-shaped bottom-up filling is preferred because V-shaped bottom-up filling can create seams. Conformal packing means that the precipitate grows from the sidewalls and bottom to the center of the feature. The most challenging features usually have very large aspect ratios (aspect ratio is the ratio of depth to width), and conformal filling typically involves a seam in the center of such features. form. Seams can be fuzzy or sharp depending on the filling mechanism. However, after annealing, any seam can create a seam void or a center void.

"Substantially void-free" means that at least 95% of the plated features or openings are void-free. Preferably, at least 98% of the plated features or openings are void-free, and most preferably, all of the plated features or openings are void-free.

"Substantially seamless" means that at least 95% of the plated features or openings are seamless. Preferably, at least 98% of the plated features or openings are seamless, and most preferably all of the plated features or openings are seamless.

To carry out the electrodeposition method, a metallized substrate, an anode, an aqueous electrolytic composition, a positive terminal in electrical communication with the anode and a negative terminal in electrical communication with the metallized substrate; An electrolytic circuit is formed, including a power source comprising: Preferably, the metallized substrate is immersed in the electrolytic composition. Electrolytic current is supplied from a power source within the circuit to the electrolytic composition, thereby depositing cobalt on the metallized substrate.

The electrodeposition process preferably involves a bath temperature ranging from about 5°C to about 80°C, more preferably about 20°C to about 50°C, most preferably about room temperature, and about 0.01 to about 20 mA/cm ² , Preferably it is carried out at a current density in the range of about 0.3 to about 10 mA/cm ² . If desired, the current may be pulsed, which may result in improved uniformity of the deposit. On/off pulses and reverse pulses may be used.

Typically, electroplating baths are agitated during use. Any suitable agitation method may be used with the processes described herein, including sparging with air or inert gas, agitation of the workpiece, impingement collection, and the like. Additionally, the workpiece may be rotated in the electroplating solution. Alternatively, instead of immersing the workpiece in the electroplating solution, the workpiece may be contacted with the electroplating solution by pumping, spraying, or other means known to those skilled in the art.

The present invention also generally relates to screening methods for evaluating the suitability of bath additives in cobalt electroplating compositions.

As mentioned above, it is important to have a smooth topology for excess thickness because rough or uneven surfaces can cause defects due to chemical mechanical polishing (CMP). Additionally, rough or uneven surfaces can be particularly problematic due to differences in over-the-board (OB) thickness between features and non-feature areas, which is typically expected due to dynamic plating. These differences in excess thickness may be further enhanced depending in part on the density of the features.

Therefore, in order to minimize excessive thickness variations between feature and non-feature areas of the substrate to determine their suitability in producing cobalt deposits, latent cobalt electroplating compositions It would be desirable to have reliable screening methods for evaluating products and the additives contained therein.

As described herein, the inventors of the present invention have surprisingly discovered that cyclic voltammetry can be used to evaluate cobalt electroplating compositions containing various bath additives. discovered. The present invention contemplates an electrochemical screening method using cyclic voltammetry to evaluate cobalt electroplating compositions.

In cyclic voltammetry scans, forward scans represent non-featured areas of the plating, and backward scans represent feature areas of the plating. The potential difference between the forward scan and the backward scan is referred to herein as a "hysteresis loop." A larger hysteresis loop means a larger potential difference and vice versa. Furthermore, the dimensions of the hysteresis loop can vary with current density.

The inventors of the present invention have discovered that cobalt electroplating compositions that exhibit small hysteresis loops at high current densities positively affect overthickness topography. Hysteresis loops at high current densities in OBs are generally at least about 5 mA/cm ² and can adversely affect OB topology. The inventors of the present invention have discovered that a clear correlation exists between the hysteresis loops at high current densities and OB topologies. It was found that larger hysteresis loops are unfavorable for CMP as they can lead to differences in OB thickness between features and non-feature regions, which can cause defects in CMP. .

In one embodiment, the high current density hysteresis loop is less than 50 mV, preferably less than 40 mV, more preferably less than 30 mV, more preferably less than 20 mV, most preferably less than 10 mV. High current density refers to a range of at least about 5 mA/cm ² , preferably a range of about 5 to about 10 mA/cm ² , more preferably a range of about 7.5 to about 8.5 mA/cm ² , most preferably a range of about This means a current density of 8 mA/cm ² . In one embodiment, a hysteresis loop of less than about 50 mV, preferably less than about 40 mV, more preferably less than about 30 mV is desired at a current density of about 8 mA/ ^cm2 .

It was also observed that most of the chemistries for good gap filling performance have hysteresis loops at low current densities, but there is no clear correlation with larger hysteresis loops being able to provide better gap filling. ing.

Excessive thickness variation between feature and non-feature areas is highly desirable. In one embodiment, the excess thickness change is controlled to be less than 50 nm, more preferably less than 40 nm, or less than 35 nm, or less than 30 nm, or less than 20 nm, or less than 10 nm.
Plating experiment:
Temperature - room temperature waveform:
OCP was followed by a gradient current of 0.5-2 mA/ ^cm2 for the amount of time required for complete via filling, then 8 mA/ ^cm2 for the time required to obtain the desired excess thickness. Gradient current.
Comparative Example 1

An aqueous cobalt electroplating bath was prepared containing 2.95 g/L Co ²⁺ ions, 30 g/L boric acid, and sulfuric acid to adjust the pH to 2.75. Ethoxylated propargyl alcohol (40 mg/L) and an amine polyol with a MW of about 5000 (50 mg/L) are also added to the composition.

Substrates with submicron features are placed in the bath and electroplated from 0.5 to 8 mA/cm ² to provide seamless filling with overload. A cyclic voltammetry scan was performed using the above waveform to measure the hysteresis loop.

The OB thickness difference (uniformity) and CV hysteresis loop measurements are listed in Table 3, and the film impurities are listed in Table 4.
Example 2

An aqueous cobalt electroplating bath was prepared containing 2.95 g/L Co ²⁺ ions, 30 g/L boric acid, and sulfuric acid to adjust the pH to 2.75. Compound I (see Table 2, x=3.6) was added to the bath at a concentration of 240 mg/L.

Substrates with submicron features are placed in the bath and electroplated from 0.5 to 8 mA/cm ² to provide seamless filling with overload. A cyclic voltammetry scan was performed using the above waveform to measure the hysteresis loop.

The OB thickness difference (uniformity) and CV hysteresis loop measurements are listed in Table 3, and the film impurities are listed in Table 4.
Example 3

An aqueous cobalt electroplating bath was prepared containing 2.95 g/L Co ²⁺ ions, 30 g/L boric acid, and sulfuric acid to adjust the pH to 2.75. Compound III (see Table 2, x=3.6, and y=1) was added to the bath at a concentration of 240 mg/L.

Substrates with submicron features are placed in the bath and electroplated from 0.5 to 8 mA/cm ² to provide seamless filling with overload. A cyclic voltammetry scan was performed using the above waveform to measure the hysteresis loop.

The OB thickness difference (uniformity) and CV hysteresis loop measurements are listed in Table 3, and the film impurities are listed in Table 4.

FIG. 1 shows cyclic voltammetry (CV) scans from +10 mV vs. OCP to −2 V vs. RE for Comparative Example 1, Example 2, and Example 3 at 100 rpm and 2 mV/s.

As shown in FIG. 1, the hysteresis loop is the potential difference between forward scan and backward scan at the same current density. At low current densities, the hysteresis loop represents the gap filling ability, and at high current densities, the hysteresis loop represents the OB ISO/density uniformity performance. The hysteresis loop curve of FIG. 1 is described in Table 3. Note that 2 mA/ ^cm2 represents a low current density and 8 mA/ ^cm2 represents a high current density.

Additionally, in Table 3, the last column describes the difference in OB thickness between feature and non-feature regions using the internal test structure described herein. This internal test structure is designated as "454." The feature region of the test structure has 0.1 μm opening trenches with 0.1 μm spacing and is considered a dense feature. A non-feature area is an area adjacent to a feature area. The OB thickness difference may represent OB ISO/density uniformity performance. Smaller difference means better OB ISO/density uniformity performance. In Table 3, it is clear that Comparative Example 1 exhibits a large hysteresis loop at 8 mA/ ^cm2 and has a larger difference in OB thickness than Examples 2 and 3, which have a small hysteresis loop at 8 mA/ ^cm2 . It is.

It is believed that the hysteresis loop at low current densities may represent the gap filling ability. In other words, if two chemistries have similar hysteresis loops at low current densities, they should be able to deliver similar gap-filling performance. In Table 3, it is clear that Comparative Example 1, Example 2, and Example 3 all have similar hysteresis loops at low current density of 2 mA/ ^cm2 and similar gap filling performance in TEM or STEM. It is.

The idea of hysteresis loop theory is to develop chemistries that have large hysteresis loops at low current densities and small hysteresis loops at high current densities to simultaneously achieve both good gap filling and OB ISO/density uniformity. It is to have.

Table 4 summarizes the SIMS (Secondary Ion Mass Spectrometry) data and shows the impurity levels in the precipitate. The goal of this study is to achieve the lowest possible amount of impurities. Ideally, it would be the same as or less than VMS only. With the unique molecular structure of Example 2, it may have far fewer impurities than other chemicals and still deliver adequate gap-filling performance.

Based on that, it can be seen that the hysteresis loop at 8 mA/ ^cm2 is small, where the OB is plated and the difference in OB thickness between the feature and non-feature areas can be small.

This observation can also be supported by theory. Kinetic plating is expected on feature areas. This means that feature regions have a faster precipitate growth rate than non-feature regions. The adsorbed acetylene inhibitor on the feature area surface is also expected to be less than the non-feature area due to the fresher precipitate surface. In the forward scan portion, the potentials are scanned from low to high, and in the reverse scan portion, the potentials are scanned from high to low. Therefore, fresher precipitate surfaces are expected to form during the forward scan. As a result, in a CV scan, forward scans represent non-feature area plating and backward scans represent feature area plating.

In a CV curve, if the forward scan is more cathodic than the backward scan at similar current density, this indicates that it takes more energy to grow in the non-feature region than in the feature region. In other words, the growth of precipitates on feature regions may be enhanced, causing a difference in precipitate thickness between features and non-feature regions. This is the fundamental reason to minimize the hysteresis loop at the current density at which the OB is plated, and is important to reduce the difference in deposit thickness between feature and non-feature areas. It is.

Thus, it can be seen that the invention described herein provides improved methods and compositions for cobalt-based metallization of interconnect features in semiconductor substrates. Additionally, the invention described herein provides screening methods for evaluating and identifying suitable bath additives to achieve desired results.

Finally, the following claims claim to cover all general and specific features of the invention described herein, as well as all the scope of the invention that may lie therebetween as a matter of language. It should also be understood that it is intended to cover the description of

Claims (19)

a cobalt deposit electroplated on the surface of the substrate, the cobalt deposit electroplated on the surface of the substrate including submicron recessed features and non-feature areas, the cobalt deposit comprising: Showing horizontal, seamless bottom-up filling in submicron recessed features and plating on non-feature areas, with an excess thickness change of less than 75 nm between submicron recessed features and non-feature areas. ,
said electroplated cobalt deposit exhibits a hysteresis loop of less than 50 mV as measured potentiodynamically in a cyclic voltammetry scan at a current density in the range of 5 to 10 mA/cm ² , and said electroplated cobalt deposit exhibits a hysteresis loop of less than 90 mV measured potentiodynamically in a cyclic voltammetry scan at current densities in the range of less than 1-5 mA/ ^cm2 ;
The electroplated cobalt deposit is
(1) A cobalt ion source,
(2) boric acid;
(3) a pH adjuster;
(4) An organic additive comprising a reaction product of alkoxylated propargyl alcohol or propargyl alcohol and a second component, wherein the second component is glycidol, propylene oxide, glycidol and propylene oxide, or propylene organic additives that are glycols and glycidol;
electroplating by immersing the substrate in a cobalt electrolyte containing;
the cobalt electrolyte is at least substantially free of accelerator;
An electroplated cobalt deposit on the surface of the substrate , wherein the concentration of divalent sulfur in the cobalt electrolyte is 1 mg/L or less . The electroplated cobalt deposit of claim 1, wherein the excess thickness variation between submicron concave features and non-feature areas of the electroplated cobalt deposit exhibits less than 50 nm. 5-10mA/cm ² 2. The electroplated cobalt deposit of claim 1, wherein the potentiodynamically measured hysteresis loop in the cyclic voltammetry scan is less than 40 mV at the current density in the range of . The current density at which the hysteresis loop is measured is 7.5 to 8.5 mA/cm. ² or the current density at which the hysteresis loop is measured is 8 mA/cm ² The electroplated cobalt deposit of claim 1. The current density at which the hysteresis loop is measured is 1.5 to 2.5 mA/cm. ² 2. The electroplated cobalt deposit of claim 1, wherein the electroplated cobalt deposit is in the range of . The cobalt precipitate is
7.5-8.5mA/cm ² 6. The electroplated cobalt deposit of claim 5 exhibiting a hysteresis loop of less than 50 mV, measured potentiodynamically in a cyclic voltammetry scan, at current densities in the range of . 2. The electroplated cobalt deposit of claim 1, wherein in the cyclic voltammetry scan, forward scans represent non-feature area plating and backward scans represent feature area plating of the cobalt deposit. 8. The electroplated cobalt deposit of claim 7, wherein the cobalt electrolyte is at least substantially free of depolarizer. A method of electrodepositing cobalt on a substrate, wherein a cobalt deposit is electroplated onto a surface of the substrate including submicron recessed features and non-feature areas, the cobalt deposit comprising submicron recessed features and non-feature areas. horizontal, seamless bottom-up filling and plating on non-feature areas, wherein the excess thickness variation between the submicron concave features and the non-feature areas is less than 75 nm;
a) receiving the substrate in an electroplating chamber;
b) immersing the substrate in a cobalt electrolyte, the cobalt electrolyte comprising:
(1) A cobalt ion source,
(2) boric acid;
(3) a pH adjuster;
(4) immersing the substrate in a cobalt electrolyte comprising an organic additive, the organic additive comprising an inhibitor;
c) electroplating cobalt into the submicron recessed features and onto the non-feature areas for a period of time under conditions that achieve horizontal, seamless bottom-up filling in the submicron recessed features; plating on the non-feature area;
The cobalt precipitate is 5 to 10 mA/cm ² and the electroplated cobalt deposit exhibits a hysteresis loop of less than 50 mV measured potentiodynamically in a cyclic voltammetry scan at a current density in the range of 1 to 5 mA/cm. ² A method exhibiting a potentiodynamically measured hysteresis loop of less than 90 mV in a cyclic voltammetry scan at current densities in the range of less than 90 mV. 10. The method of claim 9, wherein the substrate has a cobalt seed layer disposed thereon, and the cobalt is electrodeposited onto the cobalt seed layer. 10. The method of claim 9, wherein the electroplated cobalt deposit exhibits an excess thickness change of less than 50 nm between submicron concave features and non-feature areas. 5-10mA/cm ² 10. The method of claim 9, wherein the potentiodynamically measured hysteresis loop in the cyclic voltammetry scan is less than 40 mV at the current density in the range of . The current density at which the hysteresis loop is measured is 7.5 to 8.5 mA/cm. ² 10. The method of claim 9, wherein: The current density at which the hysteresis loop is measured is 1.5 to 2.5 mA/cm. ² 10. The method of claim 9, wherein: The cobalt precipitate is 7.5 to 8.5 mA/cm ² 15. The method of claim 14, which exhibits a hysteresis loop of less than 50 mV measured potentiodynamically in a cyclic voltammetry scan at current densities in the range of . 10. The method of claim 9, wherein in the cyclic voltammetry scan, forward scans represent non-feature area plating and backward scans represent feature area plating of the cobalt precipitates. 10. The method of claim 9, wherein the cobalt electrolyte is at least substantially free of promoters or at least substantially free of depolarizers. The organic additive comprises an acetylene inhibitor, and the acetylene inhibitor comprises an alkoxylated propargyl alcohol or a reaction product of propargyl alcohol with glycidol, propylene oxide, glycidol and propylene oxide, or propylene glycol and glycidol. The method according to item 9. A method of making a cobalt deposit electroplated on a surface of a substrate, the cobalt deposit comprising submicron recessed features and non-feature areas, the cobalt deposit comprising: , showing horizontal, seamless bottom-up filling in submicron concave features and plating on non-feature areas, with an excess thickness change of less than 75 nm between submicron concave features and non-feature areas. can be,
The manufacturing method includes:
(1) A cobalt ion source,
(2) boric acid;
(3) a pH adjuster;
(4) An organic additive comprising an alkoxylated propargyl alcohol or a reaction product of propargyl alcohol and a second component, wherein the second component is glycidol, propylene oxide, glycidol and propylene oxide, or propylene glycol. and an organic additive which is glycidol;
electroplating by immersing the substrate in a cobalt electrolyte containing
the cobalt electrolyte is at least substantially free of accelerator;
The concentration of divalent sulfur in the cobalt electrolyte is 1 mg/L or less,
The cobalt precipitate is 5 to 10 mA/cm ² and the electroplated cobalt deposit exhibits a hysteresis loop of less than 50 mV measured potentiodynamically in a cyclic voltammetry scan at a current density in the range of 1 to 5 mA/cm. ² exhibiting a hysteresis loop of less than 90 mV measured potentiodynamically in a cyclic voltammetry scan at current densities in the range of less than
A method for producing electroplated cobalt deposits on the surface of a substrate.